Proc. Nat. Acad. Sci. USA Vol. 70, No. 5, pp. 1608-1612, May 1973

A Bound Form of Silicon in Glycosaminoglycans and Polyuronides (polysaccharide matrix/connective tissue) KLAUS SCHWARZ Laboratory of Experimental Metabolic Diseases, Veterans Administration Hospital, Long Beach, California 90801; and Department of Biological Chemistry, School of Medicine, University of California at Los Angeles, Calif. 90024 Communicated by P. D. Boyer, March 19, 1973

ABSTRACT Silicon was found to be a constituent of notably hyaluronic acid, chondroitin 4-sulfate, dermatan certain glycosaminoglycans and polyuronides, where it sulfate, and heparan sulfate, contain relatively large amounts occurs firmly bound to the polysaccharide matrix. 330-554 ppm of bound Si were detected in purified hyaluronic acid of the element, not as free silicate ions or silicic acid but as a from umbilical cord, chondroitin 4-sulfate, dermatan firmly bound component of the polysaccharide matrix. High sulfate, and heparan sulfate. These amounts correspond to amounts of bound Si were also detected in two polyuronides, 1 atom of Si per 50,000-85,000 molecular weight or 130-280 pectin and alginic acid. Stability tests and enzymatic studies repeating units. 57-191 ppm occur in chondroitin 6-sulfate, evidence for the assumption that the Si is linked heparin, and keratan sulfate-2 from cartilage, while provide hyaluronic acids from vitreous humor and keratan sulfate- covalently to the polysaccharides in question, possibly as an 1 from cornea were Si-free. Large amounts of bound Si are ether (or ester-like) derivative of silicic acid, i.e., a silanolate. also present in pectin (2580 ppm) and alginic acid (451 ppm). The bound Si is not dialyzable, does not react with MATERIALS AND METHODS ammonium molybdate, is not liberated by autoclaving or Eight of the mucopolysaccharides studied were standard 8 M urea, and is stable against weak alkali and acid. Strong alkali and acid hydrolyze the Si-polysaccharide bond. reference samples obtained from Dr. M. B. Mathews, Depart- Free, direct-reacting, dialyzable silicate is obtained. ment of Pediatrics, University of Chicago, Ill. These samples Enzymatic hydrolysis of hyaluronic acid or pectin does not were prepared under a grant from the National Heart Insti- liberate silicic acid, but leads to products of low molecular tute to serve as reference standards for research on connective weight still containing Si in bound form. It is concluded see ref. The that Si is present as a silanolate, i.e., an ether (or ester- tissue polysaccharides. (For details 14.) following like) derivative of silicic acid, and that R1-O-Si-O-R2 materials were analyzed (see Table 1): hyaluronic acid or R1-O--Si-O-Si-O-R2 bridges play a role in the (standard reference sample), from human umbilical cords, structural organization of glycosaminoglycans and poly- molecular weight 230,000; hyaluronic acid, grade 1, from uronides. Thus, Si may function as a biological cross- human umbilical cord, Sigma Chemical Corp.; hyaluronic linking agent and contribute to architecture and resili- ence of connective tissue. acid from bovine-vitreous humor, P. L. Biochemicals, Inc.; chondroitifi 4-sulfate (standard reference sample) from noto- Silicon (Si) is essential for growth and general development cord of rock sturgeon (Acipenser fulvescens) 98% Cu4 sulfated (1). In all-plastic isolators that exclude the element from the disaccharide, molecular weight 12,000; chondroitin 4-sulfate environment (2, 3), growth of rats is reduced by 30-35% from rat costal cartilage, courtesy of Dr. M. B. Mathews; when Si-deficient aminoacid diets are fed, and bone de- chondroitin 6-sulfate (standard reference sample) from human formations develop. Dietary supplements of silicate prevent umbilical cord (15), 80% C-6 sulfated, 10% Ca4 sulfated, and these symptoms. Similar findings have been reported for 10% unsulfated disaccharide, molecular weight 40,000; chicks (4). Si, moreover, is needed in rats for normal pigment chondroitin 6-sulfate from human cartilage, courtesy of Dr. formation in the enamel of incisors (5). M. B. Mathews; chondroitin 6-sulfate from sturgeon cartilage, Even though it has been known for years that Si occurs in courtesy of Dr. M. B. Mathews; dermatan sulfate (standard small, varying amounts in all animals, very little concrete reference sample) from hog mucosal tissue byproducts of evidence exists about its functional significance and bio- heparin isolation, molecular weight 27,000; heparan sulfate chemical behavior (6, 7). In mammals, Si occurs most abun- (standard reference sample) from beef lung byproducts, dantly in connective tissue and related structures. While calcium salt, assumed molecular weight 15,000-17,000; amounts in blood and parenchymal organs are relatively low, heparin (standard reference sample) from hog mucosal tissues, the Si content of skin, cartilage, ligaments, and other tissues anticoagulant activity 180 U.S.P. international units/mg, of mainly mesodermal origin frequently exceeds 100 ,g/g dry molecular weight 11,000; keratan sulfate-1 (standard reference wt (8)*. sample) from bovine-corneal tissue, molecular weight 16,000; In extensive experiments with cartilage and bovine-nasal keratan sulfate-2 (standard reference sample) from human septum we found Si to be strongly bound -to the organic costal cartilage, heterogenous, containing portions of ma- matrix of these tissues. During attempts to identify the bind- terial incompletely substituted by sulfate; pectin (pectinic ing site of Si we discovered that certain glycosaminoglyeans, acid), purified, from citrus fruit, Nutritional Biochemicals Corp., molecular weight 100,000-200,000; alginate from Laminaria digitata, courtesy of Dr. Arne Haug, Institute for * Reported Si amounts vary greatly, depending on the analytical methods used. A large number of results (9, 10) obtained before Marine Biochemistry, Trondheim, Norway, containing of suitable methods (11, 12) and the advent of about 60o D-mannuronic and 40% Laminaria-guluronic acid; development of Dr. plastic laboratory ware are much higher than those found more alginate, prepared from Laminaria digitata, courtesy recently (8, 13). Many of these older data must be discarded. Arne Haug, containing about 40% D-mannuronic acid and 1608 Downloaded by guest on September 26, 2021 Pw.'Nat. Acad. Sci. USA 70 (1973) Bound Si in Glycosaminoglycans and Polyuronides 1609

60% Laminaria-guluronic acid, average molecular weight TABLE 1. Free, total, and bound Si in glycosaminoglycans, 200,000; glycogen, purified, from rabbit liver, Nutritional polyuronides, and some glycans Biochemicals Corp., average molecular weight 1,000,000, total ash 0.13%; starch, purified, from corn, General Bio- Si (p4g/g) chemicals, Inc.; Dextran (Clinical Grade), Nutritional Bio- Bound chemicals Corp., molecular weight 200,000-300,000 (median (total minus 254,000), prepared from culture medium of Leuconostoc Substance and source Free Total free) mesenteroides (NRRL-B512), 1,6-linked with slight trace of 1,3-linked glucoside, ash; inulin, purified, from Dahlia Glycosaminoglycans 0.02% Hyaluronic acid tubers, Sigma. (a) Human umbilical cord 25 354 329 Plastic laboratory ware was used in all procedures, except (b) Human umbilical cord 1533 1892 359 for pH measurements done with a miniature combination (c) Bovine vitreous humor 980 949 glass electrode. Samples were kept covered during practically Chondroitin 4-sulfate all manipulations to avoid Si contamination from dust. (d) Notocord of rock sturgeon 44 598 554 Si was determined by the technique devised by King and (e) Rat costal cartilage 30 361 331 collaborators (11), as modified by McGavac et al. (12, 13). Chondroitin 6-sulfate The method is based on the formation of silicomolybdate (f) Human umbilical cord 45 123 78 (SiO2. 12Mo02) and the production of a blue color upon re- (g) Human cartilage 36 227 191 duction. Interference by phosphate ions was not a problem in (h) Sturgeon cartilage 64 121 57 these ana ses. In addition to colorimetric Si determinations, Dermatan sulfate (i) Hog mucosal tissue 46 548 502 several materials were subjected to emission spectroscopy, by Heparan sulfate Mr. George Alexander and staff, Laboratory of Nuclear (j) Beef lung 39 466 427 Medicine and Radiation Biology, University of California at Heparin Los Angeles. (k) Hog mucosal tissue 33 175 142 In the colorimetric assay, 1 ,ug of Si produced an absorbance Keratan sulfate-1 of 0.167; the detection limit was 0.03 ,g of Si. Complete re- (1) Bovine cornea 31 37 agent blanks were applied side-by-side with the analyzed Keratan sulfate-2 specimen. Whenever samples were subjected to procedures (m) Human costal cartilage 37 105 68 such as dialysis, autoclaving, or hydrolysis, blanks were Polyuronides simultaneously treated in identical fashion. The following Pectin chemicals were used: ammonium molybdate, (NH4)6Mo7024 - - (n) Citrus fruit 5 2586 2581 4H20, sodium sulfite, ferric chloride, sodium acetate, an- Alginic acid hydrous powder (all analyzed reagents); and 1,2,4-amino- (o) Horsetail kelp 43 naphthol-sulfuric acid (all from J. T. Baker Chemical Co.); (p) Horsetail kelp 5 456 451 sulfuric acid, 95-97%, A.C.S. reagent (Dupont); sodium Polyglycans metabisulfite, reagent (Matheson, Coleman & Bell); sodium Glycogen carbonate, anhydrous powder, A.C.S. analytical reagent, and (q) Rabbit liver 8 34 26 hydrochloric acid, analyzed reagent, U.S.P. (Mallinckrodt); Starch congo red, certified (Allied Chemical). (r) Corn - 22 The colorimetric method was used to determine the amount Dextran of free, directly reacting silicate and silicic acid, and to (s) Leuconostoc mesenteroides 19 22 determine the amount of total Si after sodium carbonate fusion Inulin in platinum crucibles. In some cases, indirect reacting Si was (t) Dahlia tubers 15 29 14 measured after treatment of various polysaccharides with dilute sodium hydroxide at 1000. occasions free silicate was added in microgram quantities to Details of experiments to liberate bound Si by dialysis, glycosaminoglycan or polyuronide solutions. In experiments autoclaving, 8 M urea, alkali, or acid are described in Table 2. done under various conditions over extended periods of time, For enzymatic hydrolysis of hyaluronic acid, hyaluronidase no evidence could be found for spontaneous formation of from bovine testes was used (Sigma, 348 units/mg). De- bound Si. polymerization of hyaluronic acid was measured by the Bound Si in Glycosaminoglycans and Polyuronides. Com- turbidimetric assay (16) with fraction V bovine albumin parison of the free, directly reacting silicate to total Si de- (Sigma). For the enzymatic breakdown of pectin, pectinase termined after carbonate fusion reveals that large amounts of from Aspergillus niger was used (Sigma, approximately 1 bound Si occur in certain acid mucopolysaccharides and poly- unit/mg solid). Pectinase activity was followed by iodometric uronides (Table 1). 13 Samples of acid mucopolysaccharides titration of reducing end groups (17, 18). and three polyuronides were studied. In addition, four com- monly occurring polyglycans were analyzed. Of all 20 samples, RESULTS only two contained large amounts of free, directly reacting Determination of Free, Unbound Silicic Acid. Free, unbound silicate ions (1533 and 980 ,ug/g); both were commercial silicic acid in mucopolysaccharides can be directly determined preparations of hyaluronic acid. All other substances showed by colorimetric analysis since, within reasonable limits, the negligible amounts of free silicate. 330-554 ppm of bound Si presence of mucopolysaccharides and polyuronides does not were detected in hyaluronic acid prepared from umbilical cord, interfere with the colorimetric Si determination. On numerous chondroitin 4-sulfate, dermatan sulfate, and heparan sulfate. Downloaded by guest on September 26, 2021 1610 Biochemistry: Schwarz Proc. Nat. Acad. Sci. USA 70 (1973)

Smaller amounts of bound Si (57-191 ppm) occurred in chon- duced practically identical data, for example in samples d, k, droitin 6-sulfate, heparin, and keratan sulfate-2 from cartilage. q, and r of Table 1 (600 compared to 598, 180 compared to 175, Hyaluronic acid from vitreous humor and keratan sulfate-1 33 compared to 34, and 24 compared to 22 Ag/g). In other in- from cornea were practically free of bound Si. stances larger differences appeared, e.g., in samples a, j, m, p, The highest amount of bound Si was encountered in purified and t (482 compared to 354, 580 compared to 466, 80 com- pectin from citrus fruit, which contained over 2500 ppm by pared to 105, 360 compared to 456, and 12 compared to 29 colorimetric analysis. Two samples of alginic acid from horse- Jug/g). Where one contained 456 tail kelp differed in their Si content. Experiments to Liberate Bound Si were done mainly with ppm, an amount comparable to that found in acid mucopoly- and estimation of free hyaluronic acid (sample a, standard reference sample) saccharides, the other was low in Si. A direct obtained with these two ma- silicate in the latter material was not possible since a pre- pectin (sample n). The results molybdate- terials were analogous, but the Si in pectin was less readily cipitate formed upon addition of the ammonium or acid (Table 2). While silicic acid ions sulfuric acid reagent. No such problem existed in direct liberated by alkali for were dialyzable against water or 0.1 N NaOH when added to analyses of the other samples tested, except cornstarch. or polyuronides, only a four polyglycans, glycogen, starch, dextran, and inulin, solutions of mucopolysaccharides The small fraction of the bound Si was removed by extended contained only minute amounts of Si, most probably as corre- under conditions that dialysis (Exp. I, Table 2). The amount dialyzable silicate. All had been manufactured liberated by prolonged standing in 0.1 N avoided exposure to strong alkali or acid. This consideration sponded to that Si and the carbohydrate NaOH at room temperature (240). Neither autoclaving is important since the bond between (Exp. III), which to alkali and acid. (Exp. II) nor treatment with 8 M urea matrix is sensitive could break hydrogen bonds, liberated the bound Si. A small Comparison of Results of Colorimetric and Emission Spectro- quantity of free silicic acid appearing after autoclaving of scopic Si Analysis. The results of Si determinations by dc arc hyaluronic acid may have been due to hydrolysis, since the emission spectroscopy in general confirmed those of the experiment was performed at pH 5.7. colorimetric method. In some cases, the two methods pro- Alkaline hydrolysis broke the bond between Si and the polysaccharide molecule (Exps. IV and V). Free silicate was formed that reacted directly with ammonium molybdate TABLE 2. Stability of Si binding in hyaluronic acid (HA)a and was dialyzable. At room temperature, hyaluronic acid and pectinb lost only 12% of the Si after 60 hr in 0.1 N NaOH, but in 1 N NaOH was liberated within 2 hr (Exp. IV). The Si in Free Liberated 88% Si Si (%) pectin remained bound under these conditions, and even (Cg/g) after 18 hr at pH 13.5, only 15% of the bound Si had been Treatment HA Pectin HA Pectin released. At 1000 (Exp. V), 0.1 N NaOH liberated almost all 2 but from Control 25 20 - of the Si from hyaluronic acid within hr, only 29% (I) Dialysis, 0.1 N NaOHO d pectin. 2-hr treatment with 1 N NaOH at 1000 liberated 77%. Dialysate, 70 hr, at 40 42 25 5 0 Prolonged alkaline hydrolysis was needed to completely re- (II) Autoclavinge cover the bound Si from pectin as silicate ion. 15 lb pressure, 1 hr 49 41 7 <1 At pH 3, the bond between Si and polysaccharide matrix (III)Urea, 8 Mf was stable over extended periods of time, but treatment with 20 hr at 40 22 43 0 <1 0.2 N HC1 (pH 1.2) liberated a large portion of the bound Si lhrat600 23 36 0 <1 (Exp. VI). In acid, as in alkali, hyaluronic acid released its (IV)Alkali at room temperature (240) bound Si more readily than pectin. Within 2 hr, 45% of the Si pH 12.4, 0.1 N NaOH, 60 hr 60 28 12 <1 from hyaluronic acid was found as free, directly-reacting pH 13.5, 1 N NaOH, 2 hr 316 40 88 <1 100 15 silicic acid, and after 18 hr, 58% had been removed. Much of pH 13.5, 1 N NaOH, 18 hr 399 392 Si in the (V)Alkali at 1000 it precipitated out as a sediment of silica. Tests for pH 12.5, 0.1 N NaOH, 2 hr 313 460 95 29 supernatant by sodium carbonate fusion showed that the pH 13.4, 1 N NaOH, 2 hr - 1960 - 77 conversion was incomplete. Under these conditions, pectin (VI)Acid at room temperature (24°) released only 9% of its bound Si; heating was required for pH 3.0, 18 hr 20 22 0 0 complete liberation (details to be published elsewhere). pH 1.2, 0.2 N HCl, 2 hr 147 32 45 <1 pH 1.2, 0.2 N HCl, 18 hr 1909 242 58 9 Effect of Enzymatic Hydrolysis on Bound Si. Studies with hyaluronidase and pectinase showed that the enzymatic a Tests performed on solutions containing, per ml, 1 mg of breakdown of the polysaccharide structure does not convert hyaluronic acid, sample a (329 jug bound Si per g). b Tests per- the bound Si into free, directly-reacting silicic acid (Table 3). formed on solutions containing, per ml, 5 mg of pectin, sample n Instead, breakdown products of low molecular weight were (2581 u~g bound Si per g). 010 ml of solution dialyzed against 100 obtained to which the Si is firmly attached. Under the condi- ml of 0.1 N NaOH for 5 hr, followed by a second dialysis for 65 hr. tions applied (Exp. I), hyaluronidase rapidly depolymerized d Experiments with nearly identical results done with heparan the hyaluronic acid. After 60 min, turbidity readings sulfate (sample j) and heparin (k). e Hyaluronic acid autoclaved amounted to only 0.7% of those of the initial reaction mixture at pH 5.7, pectin at pH 7.0. f 5 ml of hyaluronic acid solution, No of silicate ions was detected. the 8 M urea read a at zero time. release Hy- 2.4 g of urea. Measurements of Si in against 12.5 for 2 hr at 1000 M urea. £ as SiO2- drolysis of the incubation mixture at pH standard curve done in 8 Partially separated of xH20, dissolved in 1 N NaOH for analysis. Determination of Si was used to release bound Si (indirect Si). The equivalent in the supernatant showed 157 ,ug still in bound form. 300 jAg of indirect Si was found per g in the reaction mixture. Downloaded by guest on September 26, 2021 Proc. Nat. Acad. Sci. USA 70 (1973) Bound Si in Glycosaminoglycans and Polyuronides 1611

Untreated hyaluronic acid produced 319 /Ag of indirect Si TABLE 3. Effect of enzymatic hydrolysis on bound Si under the same conditions. In a subsequent experiment (Exp. II), an attempt was made to break the hyaluronic acid Indirect down to tetrasaccharide units, as described by Ludowieg et al. Free (19). Again, free silicate was not detectable in the incubation Si (,g/g) Sis (ug/g) Liberated (%) mixture, but a large amount of bound, indirect Si was found. It was dialyzable against water without decomposition. (I)Hyaluronic acid, 60 After nearly complete enzymatic hydrolysis of pectin by min incubationb pectinase (Exp. III) only of the Si bound to the Before 33 329 6% initially Aftero 33 300 0 pectin matrix was found in free form. The main portion was (II)Hyaluronic acid, 45 hr present as bound, indirect Si. As shown above, however, the incubationd standard treatment for 2 hr at pH 12.5 was inadequate to Before 32 322 completely liberate the bound silicic acid present in pectin. After e 321 0 The enzymatic breakdown products of pectin containing the (III)Pectin, 20 hr bound, indirect reacting form of Si were dialyzable. incubationt Thus, the results of the enzymatic hydrolysis of hyaluronic Before 20 2560 - acid and pectin provide further evidence that Si si covalently Afterg 179 1279h 6 bound. It appears to be an integral component of these sub- stances in their natural state. a Si liberated at pH 12.5 after heating to 1000 for 1 hr. b Mix- ture containing 0.6 mg of hyaluronic acid, sample a, and 0.1 DISCUSSION mg of hyaluronidase per ml 0.1 M sodium acetate buffer pH 5.3, The data show that firmly bound Si is a constituent of certain kept for 60 min at 37'. C Turbidity readings in depolymerization mucopolysaccharides and polyuronides, notably hyaluronic assay dropped from 1.489 for the initial reaction mixture to 0.009 chondroitin dermatan after 60 min. d Conditions as in b. Fresh enzyme added after 3 acid, 4-sulfate, sulfate, heparan sulfate, and 24 hr. e Slightly turbid. No blue color. f Mixture containing and pectinic and alginic acids. We were unable to find any 5 mg of pectin, sample n, and 4 mg of crude pectinase per ml at published reference to previous findings on the occurrence of pH 4.0, kept for 20 hr at 250. g Iodometric titration revealed that Si in such compounds. t The element appears to be present as a about 93% of galacturonic acid had been liberated. h Incomplete derivative of silicic acid. Bound Si does not react with ammo- "indirect'" analysis (see text). nium molybdate and is not dialyzable. Since it is not liberated by autoclaving or treatment with 8 M urea and is stable against dilute alkali and acid at room temperature, hydrogen also possible, however, that Si is specifically connected to bonding does not seem to play a decisive role in Si binding. certain minority sugars that may occur at interspersed inter- The findings thus lead to the conclusion that Si in vals in the polymeric glycan chain and act as tie-up sites. glycos- Bound Si may play an aminoglycans and polyuronides is covalently bound to the important role in the structural carbohydrate matrix, most likely as a silanolate, i.e., an ether organizaiion of acid mucopolysaccharides and polyuronides. (or ester-like) derivative of silicic acid and hydroxyl groups. In principle, Si could link two binding sites through a Rj-0 The between the Si atom and the Si-O-R2 group; the Si atom would carry two additional bridge carbohydrate polymer radicals. chain would consist of a Si-O-C group. This concept is Orthosilicic acid, Si(OH)4, could connect up to four strengthened by the results of the enzymatic studies. binding sites directly. An alternative exists in which the Si Acid mucopolysaccharides with large amounts of bound Si atoms of two compounds would be linked by an oxygen-bridge, are similar to each other in Si content, even though they are forming a R1-O-Si-O-Si-O-R2 group; the two Si very dissimilar with respect to molecular weights. The Si- atoms could bind up to four additional radicals by way containing acid bind 1 atom of Si of oxygen. Si has a strong tendency to form such groups mucopolysaccharides per (7, 22). This 50,000-85,000 molecular weight; this corresponds to 1 atom of tendency could lead to attachment of further Si Si per 130-280 repeating units. Calculation of the molar ratios atoms and formation of Si-O ring systems or polymers. shows that there are, per atom of Si, about 0.3 molecule of Thus, firm structural arrangements between bundles of hyaluronic acid from umbilical cord (molecular weight 230,- mucopolysaccharides, and also proteins, are possible. Such 000), 2 molecules of dermatan sulfate (molecular weight interlacing structures could hold acid mucopolysaccharides 27,000), 4 mol of chondroitin 4-sulfate (molecular weight and also proteins together in an organized fashion, contribut- 12,000), 4 mol of heparan sulfate (if average molecular weight ing to the architecture and resilience of connective tissues, is assumed to be about 16,000), and 8 mol of chondroitin 6- e.g., cartilage (23). Si functions as a biological crosslinking sulfate. It remains to be seen whether these figures are coinci- agent in connective tissue. Stereochemical questions arise dental or whether they express an innate regularity of muco- from these considerations since the substituents of the Si polysaccharide chemistry. The pectin contains 10-20 atoms of atom, like those of the carbon atom, are arranged in tetra- Si per mol (molecular weight 100,000-200,000), i.e., 1 atom of hedral geometry that is much more rigid than that in carbon Si per 10,000 molecular weight. The minimum molecular chemistry. "Minimization of motion and angle distortion for weight of pectin is about 10,000 (21). The Si could be attached nonreacting groups is an important factor in organosilicon to a suitable hydroxyl group of the uronic acid moiety. It is mechanisms" (24). With respect to individual polysaccharide molecules, bound t However, Holt postulated in 1953 that in silicosis, silicic acid Si could function in several ways: (a) It could connect differ- from the inhaled silica dust replaces normal mucopolysaccharides ent portions of the same polymeric saccharide chain, estab- in the interfibrillary spaces, thus causing the pathological changes lishing a secondary structure and controlling molecular shape. in the lung (20). (b) It could link different polysaccharide chains to each other. Downloaded by guest on September 26, 2021 1612 Biochemistry: Schwarz Proc. Nat. Acad. Sci. USA 70 (1973)

Si may contribute to the very high molecular size of glycos- and on Si-containing intermediates appear indicated. The aminoglycans and polyuronides in their natural state. (c) finding of Si as a constituent of glycosaminoglycans and Si could also link acid mucopolysaccharides to proteins. Almost polyuronides introduces a new aspect into the discussions all glycosaminoglycans, and many other polysaccharides, surrounding the chemistry of these compounds. occur in nature primarily as protein-polysaccharide com- I thank Dr. M. B. Mathews, Department of Pediatrics, pounds (25). Si could establish a bridge between glycosami- University of Chicago, Ill., for generous gifts of acid mucopoly- noglycans and collagen, or the globular protein found in saccharides; Dr. Arne Haug, Institute for Marine Biochemistry, the ground substance of connective tissue (26). Such a link Trondheim, Norway, for alginic acids; and Mr. George Alex- could exist aside from the well-documented polysaccharide- ander, Laboratory of Nuclear Medicine and Radiation Biology, UCLA, for emission spectroscopic analyses. The excellent xylosyl-serine-protein bridge and similar links found in pro- technical assistance of Mrs. Billie Ricci, Mrs. Martha Duck- teoglycan molecules (27). worth, and Mr. Dennis Kim is gratefully acknowledged. This A consistent difference is seen between samples of chon- work was supported by USPHS Grant AM 08669. droitin 4-sulfate and those of chondroitin 6-sulfate, the latter 1. Schwarz, K. & Milne, D. B. (1972) Nature 239, 333-334. containing only a fraction of the quantity of bound Si found 2. Smith, J. & Schwarz, K. (1967) J. Nutr. 93, 182-188. in the former. Low amounts of Si could either be character- 3. Schwarz, K. (1970) in Trace Element Metabolism in Animals, istic of chondroitin 6-sulfate, or they could be due to impuri- ed. Mills, C. F. (E&S Livingstone, Edinburgh), pp. 25-38. ties. In view of the great difficulties in obtaining mucopoly- 4. Carlisle, E. M. (1972) Science 178, 619-621.. 5. Milne, D. B., Schwarz, K. & Sognnaes, R. (1972) Fed. saccharides in pure form, other samples in which we found low Proc. 31, 700. amounts of bound Si may also not have been homogeneous. 6. Garson, L. & Kirchner, L. (1971) J. Pharm. Sci. 60, 1113- The data on pectin and alginic acid show that the occur- 1127. rence of bound Si is not limited to acid mucopolysaccharides of 7. Fessenden, R. & Fessenden, J. (1967) Advan. Drug Res. 4, 95-132. animal origin.: However, the preparation of these products 8. Fregert, S. (1959) Acta Dermato-Venereol. 39, suppl. 42, normally involves treatment with strong alkali or acid, often 1-92. at elevated temperatures, conditions under which Si may be 9. King, E. & Belt, T. (1938) Phys. Rev. 18, 329-365. lost. 10. Holt, P. & Yates, D. (1953) Biochem. J. 54, 300-305. Numerous have been made to extract 11. King, E., Stacy, B., Holt, P., Yates, D. & Pickles, D. attempts and identify (1955) The Analyst 80, 441-453. organic Si compounds from tissues, but in no case has a pure 12. McGavack, T., Leslie, J. & Tang Kao, K. (1962) Proc. Soc. substance been isolated and chemically characterized (6, 7). Exp. Biol. Med. 110, 215-218. Si "esters" of lipids such as cholesterol, lethicin, and choline 13. Leslie, J., Tang Kao, K. & McGavack, T. (1962) Proc. Soc. have been described after extraction of tissues by ethanol- Exp. Biol. Med. 110, 218-220. 14. Roden, L., Baker, J. R., Cifonelli, J. A. & Mathews, M. B. ether. Holt and Yates demonstrated that such products are (1972) in Methods of Enzymology, ed. Ginsburg, V. (Aca- artifacts, obtainable in vitro by treating these compounds with demic Press, New York), Vol. 28B, pp. 73-140. soluble silicic acid. They arise from micelle formation of 15. Mathews, M. B. (1966) in Methods of Enzymology, eds. inorganic oligo- or polysilicate ions with organic compounds Neufeld, E. F. & Ginsburg, V. (Academic Press, New (10). Similar derivatives have been described for carbohy- York), Vol. 8, pp. 654462. 16. Kertesz, Z. I. (1955) in Methods of Enzymology, eds. Colo- drates. A critical review of these studies, and of Si in plant wick, S. P. & Kaplan, N. 0. (Academic Press, New York), growth, is available (28). Because of the stability of the bound Vol. 1, pp. 152-164. form of Si in glycosaminoglycans and polyuronides presented 17. Owens, H. S., McCready, R. M., Sheperd, A. D., Schultz, here, micelle formation can be excluded from the interpreta- T. H., Pippen, E. L., Swenson, H. A., Miers, J. C., Erland- sen, R. F. & Maclay, W. D. (1952) Methods used at Western tion of our findings. Regional Research Laboratory for Extraction and Analysis of The metabolic processes by which Si is assimilated are little Pectic Materials (U.S. Dept. Agriculture, AIC-340). understood, even though large amounts of the element are 18. Worthington Enzyme Manual (1972) (Worthington Bio- used for structural stability by diatoms, many higher plants, chemical Corp., Freehold, New Jersey), p. 112. and two groups of animals, the Radiolaria and 19. Ludowieg, J., Vennesland, B. & Dorfman, A. (1961) J. (protozoa) Biol. Chem. 236, 333-339. sponges. Limpets (Patellacea) assimilate Si to make opal 20. Holt, P. F. & Osborne, S. G. (1953) Brit. J. Ind. Med. 10, baseplates for their teeth (29). In diatoms, Si supplementation 152-156. stimulates numerous enzyme systems and leads to an imme- 21. Kertesz, Z. I. (1963) in Comprehensive Biochemistry, eds. diate increase in amounts of DNA. Inhibitors of energy Florkin, M. & Stotz, E. H. (Elsevier, N. Y.), Vol. 5, pp. 233-245. metabolism interfere with the uptake of silicate. 22. Cotton, F. A. & Wilkinson, G. (1966) Advanced Inorganic Since Si is essential for higher animals and since it is Chemistry (Interscience Publishers, New York), pp. 468-474. apparently an integral constituent of acid mucopolysaccha- 23. Barrett, A. J. (1968) in Comprehensive Biochemistry, eds. rides and polyuronides, the question arises whether there are Florkin, M. & Stotz, E. H. (Elsevier, New York), Vol. 26B, common pathways of Si metabolism in bacteria, plants, and pp. 438-442. 24. Sommer, L. H. (1965) Stereochemistry, Mechanism and animals. With respect to mucopolysaccharides, it is theoreti- Silicon (McGraw-Hill, New York). cally possible that inorganic Si is bound after the macromolec- 25. The Chemical Physiology of Mucopolysaccharides (1968) ed. ular structure has been formed. An alternative, more plausible Quintarelli, G. (Little, Brown & Co., Boston). from stereochemical considerations, would consist of the 26. Partridge, S. M. (1968) in The Chemical Physiology of incorporation of mono- or Si deriva- Mucopolysaccharides, ed. Quintarelli, G. (Little, Brown & preformed disaccharide Co., Boston), pp. 51-62. tives during synthesis of the polysaccharide chain. Investiga- 27. Barrett, A. J. (1968) in Comprehensive Biochemistry, eds. tions on the enzymatic formation of biological Si derivatives Florkin, M. & Stotz, E. H. (Elsevier, New York), Vol. 26B, pp. 434438. 28. Lewin, J. & Reimann, B. E. F. (1969) Annu. Rev. Plant. $ In cotton, considered the purest form of naturally occurring Physiol. 20, 289-304. cellulose, 116 jg of Si were found per g. 29. Lowenstam, H. (1971) Science 171, 487490. Downloaded by guest on September 26, 2021